[bip] Blog post on bioinformatics and Python

[Andrew Dalke]

On Sep 16, 2008, at 1:12 PM, Leighton Pritchard wrote:
> Because if your aim is to build a general purpose bioinformatics  
> library, it
> might be more efficient and sensible to re-use, benefit from and  
> build upon
> the pre-existing open-source Biopython,  ...

A: Hey, we like Python and do bioinformatics.  Let's develop something
like Bioperl
J: Great!  The domain is available and Bioperl is willing to host us.
A: And if we use a BSD license then people can at the very least look
   at what we've done and use it to base new projects.
I, B, and others: here's code, and tests, and organizational support.

  ... over time ...

A: I mostly do cheminformatics these days  ...
J: I've got a family and post-doc to focus on ...
   (chorus - sounds of sad agreement)
M: I can do more with it.
P: So can I
   (chorus - sounds of happy agreement)

   ... over time ...

1: Biopython sure has a lot of parts
2: And I hear they don't quite go together
1: It surely must be easier to start a new project using fresh,  
modern ideas
2: We'll be sure not to make the same mistakes
2: Why, we won't even look at their code - that way we can avoid  
their problems.


Exaggerating for emphasis - If you are developing a reusable library,  
and don't yourself reuse components, then you will have problems  
understanding the needs of your user base.  You are someone who wants  
to write new code while your users are people who want to use  
existing tools.

(Am I roiling the pot enough?)


				Andrew
				dalke at dalkescientific.com